Background

One challenge in large animal medicine regarding prudent use of antibiotics is that clinical signs alone cannot be used to determine underlying aetiology. Examples are respiratory disease with both bacterial and viral aetiologies [1], mastitis where some cases will self-resolve [2] and lameness with both infectious and traumatic causes [3]. Among the diagnostic tools are post-mortem examinations and laboratory testing e.g. PCR, bacterial cultivation and somatic cell count [1, 2, 4], which can be laboursome and expensive. Further, these procedures may not be compatible with the need to promptly decide if antibiotic treatment is necessary when confronted with one or more sick animals [5, 6]. Consequently, veterinarians are at risk of using and prescribing antibiotics based on empirical diagnosis and without identifying the causative agent [5], which can lead to overuse of antibiotics and further increase the risk of antimicrobial resistance [7, 8].

Acute phase proteins (APPs) are key elements of the innate immune system and are a part of the acute phase response (APR). The acute phase proteins are mainly synthesized by hepatocytes and the production is cytokine dependent. Interleukin-6 (IL-6), interleukin 1 (IL-1) and tumor necrosis factor-alpha (TNF-α) are major inducers [9, 10]. Generally, the response is unspecific and connected to both infection, trauma or inflammation [11]. However, other conditions such as heat stress [12], parturition [13, 14] and intensive training [15, 16] have been associated with altered APP levels in animals.

C-reactive protein (CRP) and serum amyloid A (SAA) are used in humans to differentiate between bacterial and non-bacterial causes of disease, and to guide initiation of antibiotic treatment, thus potentially reducing unnecessary antibiotic use [17, 18]. Additionally, CRP could reduce antibiotic treatment duration in children and treatment initiation in adults without adverse negative effects [19, 20]. C-reactive protein differentiate infectious from non-infectious aetiologies which could prevent unnecessary antibiotic initiation in diseases such as systemic inflammatory response syndrome [20], respiratory disease [19, 21], diagnosis of septic arthritis and osteomyelitis [22] and unspecific fever [23, 24]. The widespread use of APPs to optimize antibiotic use in humans, as well as a recently published review on biomarkers and sepsis in animals [25], suggests that APPs could be a valuable tool in veterinary medicine.

The acute phase response is species and disease dependent. This means that the magnitude of the response and which APPs are affected differ between species and conditions [10, 11]. Therefore, knowledge of the APR in specific species and diseases is needed to establish the clinical utility of APPs for guiding antibiotic use in veterinary medicine. The aim of this mapping review was therefore to investigate existing knowledge on APPs in pigs, horses and cattle and the potential to optimize antibiotic treatment for diseases in these animals.

Search strategy

Relevant literature on APPs and diagnostic use of APPs was retrieved by searching Medline (1946 to March 2023), Google scholar and CAB Abstracts (1973–March 2023). Literature published up until March 2023 was considered. Abstracts were read, and relevant literature retrieved. Duplicates and studies that were not peer-reviewed or in English were excluded. Search terms used were: “Acute phase”, “Antimicrobial*”, “Antibiotic*”, “Cow”, “Pig” and “Horse”.

Review

Pigs

The porcine APR has been studied in experimentally and naturally infected pigs and in pigs with clinical signs of disease or injury. In growing pigs, the APR has been studied for a plethora of experimentally induced bacterial and viral diseases either alone or as co-infections. Examples are Streptococcus suis [26], porcine reproductive and respiratory syndrome (PRRS) [27, 28], Actinobacillus pleuropneumoniae (AP) [29], swine influenza and Pasteurella multocida [30, 31], Bordetella bronchiseptica [32], Pasteurella multocida [33], swine influenza and Bordetella bronchiseptica [34, 35] and swine influenza and Actinobacillus pleuropneumoniae [36]. In finishing pigs, the studies have focused on Actinobacillus pleuropneumoniae [37,38,39], Streptococcus suis [40], osteomyelitis [41], African swine fever (ASF) and Aujeszky’s disease (AD) [42]. In sows, experimentally induced infection of the mammary gland with E. coli has also been investigated [43]. Information regarding induced diseases, affected APPs and age groups has been summarized in Table 1.

Table 1 Overview of studies on acute phase protein (APP) levels in pigs for experimentally induced diseases1

The experimental inoculations revealed that all moderate and major APPs are indeed elevated in most diseases in the week post-infection. Exceptions are osteomyelitis (S. suis), intramammary inoculation of E. coli and swine influenza (H1N1), where SAA [41], haptoglobin (HP) [43] and CRP, pig major acute phase protein (Pig-MAP) and SAA [36] remained unchanged respectively. Regarding the elevation of HP in AP infection the results are equivocal with two studies finding no effect [36, 38] and two studies finding an effect [37, 39]. This difference could be affected by inoculation method: A viral infection comparing an aerosol to an intranasal inoculation showed that aerosol inoculation lead to a more severe pathology in the lower respiratory tract [44, 45] and induced a higher immune and cytokine response [45]. Furthermore, one of the studies with no effect followed the pigs for 18–24 h post infection [38] which might be too short a time period to detect a HP response, since HP have been shown to have a delayed response in pigs [46]. In pigherds several infections can occur simoultaneosly and evidence on the APP response in bacterial and co-infections suggests that being co-infected with viral and bacterial pathogens should not mask a bacterial infection [30,31,32,33,34,35]. However, in order to distinguish bacterial from viral infections a significant difference between APP levels is needed so that a cut-off level with high sensitivity and specificity can be established. The studies inducing PRRS, ASF and AD did not provide the magnitude of the increase or the mean APP levels pre- and post-infection. Therefore, approximations were made based on the illustrations provided. The ranges are included in Table 2. Minding the decreased reliability, CRP levels seemed lower in viral infections compared to bacterial infections. Additionally, CRP but not HP, followed the reversion of clinical signs in pigs treated with danofloxacin and hence CRP might be used as a clinical evaluation of treatment efficacy [39].

Table 2 Increase in acute phase proteins in experimentally induced bacterial and viral infections in pigs

In experimentally induced infections the exact time of inoculation is known, and blood samples can be taken repeatedly to study the kinetics and establish maximum APP levels. However, in practice most infections are continously circulating in the herds and pigs might be chronically infected or several days into the acute phase before diagnostics are performed. The half-life of CRP is 47 h with elevations seen 6 h post-stimulus, for SAA the half-life is 35 h and for HP it is 3–5 days [25]. In addition, studies have shown that Pig-MAP increases 2 days post-infection and reaches normal levels within 5 to 10 days post-infection [32, 36]. Therefore, if blood samples are not performed within the first days post-infection, the CRP and SAA levels migth already have decreased below physiological levels. In addition, the pigs might suffer from a mixture of infections and injuries, which might blur the response. Hence in order to conclude if APPs can be used for antibiotic guidance on farm, trials with naturally occurring diseases and clinical signs are needed. One such study investigated the correlation between HP levels and different clinical signs of disease in seven specific pathogen free (SPF) herds and four non-declared herds [47]. Specific pathogen free is a Danish health system which is used to declare the health status of a herd regarding seven specific pathogens. The targeted diseases are enzootic pneumonia, porcine pleuropneumonia, porcine reproductive and respiratory syndrome, swine dysentery, atrophic rhinitis, mange and lice [48]. The study focused on respiratory disease, atrophic rhinitis, lameness, tail or ear bite, diarrhoea, umbilical hernia and neurological disease in finishing pigs. The study found significantly increased HP levels in pigs with clinical signs of lameness and tail or ear bite (P < 0.001). However, no difference was seen in pigs with respiratory signs or umbilical hernias. Diarrhoea, acute rhinitis and neurological signs were excluded due to low prevalence. Additionally, the study found a herd effect and an interaction between the herd health status (SPF vs. non-declared) and the age of the finishing pigs (P < 0.05) [47]. An effect of herd health status on acute phase protein levels has been shown in another study [49] indicating that baseline APP levels can differ between herds with differing health status, which should be considered when interpreting the magnitude of the APR.

As can be seen from Table 3, significant increases in one or more APPs were found for all diseases. C-reactive protein and SAA were elevated in all diseases but AD, HP in all diseases, Pig-MAP was elevated in Porcine circovirus type 2 (PCV2), M. hyopneumoniae and inflammation and albumin (Alb) was significantly decreased in PRRS, PCV2 and inflammation. Interestingly, and in line with findings from experimentally induced infections, CRP was the only APP where a larger response was apparent in bacterial disease (71-fold increase), compared to viral (9 and 26-fold increase) and inflammation (38-fold). Indicating that CRP might, as seen in humans [18, 50], be a candidate for antibiotic guidance in pigs.

Table 3 Overview of acute phase protein (APP) levels in pigs for common diseases and clinical signs1

The mean APP levels of three other studies performed on suckling, growing and finishing pigs has been summarized in Table 3. In accordance with the other studies, significant differences between pigs with and without clinical signs were discovered [46, 50] although few discrepancies exist e.g. in respiratory disease which altered a non-significant [47] and a significant increase in HP [51]. This could be due to differences in aetiology or subclinical disease, which would also explain the high HP levels in the control group [47]. The third study, which investigated the correlation between swine inflammation and necrosis syndrome (SINS) score and APP levels, found that for suckling pigs HP and CRP correlated with SINS score, and for fattening pigs CRP and fibrinogen (Fb) correlated with SINS score [52]. Studies investigating the same porcine diseases are lacking, hence more research is needed to conclude which APPs can be used diagnostically for which diseases. However, Receiver-operating-characteristic curve (ROC-curve) analysis was used to establish the accuracy of using HP as an objective marker of clinical disease and an accuracy of 76–78% was found for lameness, diarrhoea and tail or ear bites and an accuracy of 67% was found for respiratory disease [51].

In sows, the majority of studies on APP investigated postpartum dysgalactia syndrome (PDS), a multifactorial syndrome causing decreased growth and increased mortality in piglets due to dysgalactia, fever, mastitis and/or metritis in affected sows [53]. However, one study investigated the CRP and HP response in vulvar discharge syndrome (VDS) and found no significant differences in sows with and without VDS [54]. The effect of reproductive disease on APP levels has been summarized in Table 4. The table contains data from sows with PDS and VDS, but only days where significant differences between groups were found, were included. Around parturition an APP dependent increase in levels occurs, with healthy sows showing lower rises than PDS sows [13, 14, 43]. For instance, a study found that HP, presented with two peaks around parturition in both healthy and PDS sows and that the HP levels were significantly elevated from day −14 to + 14 days postpartum [14]. For CRP, Pig-MAP and SAA a peak was seen postpartum in healthy and PDS sows, with levels elevated from farrowing and until 7–14 days postpartum [14]. The elevations seen in apparently healthy sows indicate that interpreting APP levels on sows  ± 14 days of farrowing should be done with care, since rises are not necessarily indicative of underlying inflammation or disease. Three out of four studies on PDS showed significant differences between healthy and PDS sows [13, 14, 55]. The test period and investigated APPs vary considerably between studies, but rises in CRP, HP, SAA and alpha 1-acid glycoprotein (AGP) could be indicative of PDS. However, to be clinically relevant, the alterations need to be apparent either before farrowing or within the first days after farrowing and be correlated to clinical signs of PDS. Serum amyloid A levels has been shown to be correlated (R-spearman—0.76) with PDS before farrowing. The study found sensitivity, specificity and kappa-value of 0.91, 0.99 and 0.92, respectively, using a threshhold of 0.63 [14]. However, no explanation of the clinical exam used as a gold standard for PDS sows, the inclusion and exclusion criteria in all groups and the sampling procedure was provided in the paper, making it difficult to interpret if the sows included were representative of parturating sows in general.

Table 4 Overview of acute phase protein (APP) levels in sows and gilts for naturally occurring diseases1

Summary—pigs

The moderate to major APPs in pigs are CRP, SAA, HP, and Pig-MAP, which all show a large response in most of the diseases or syndromes studied. An acute phase response was elicited by all experimental and naturally occurring bacterial and viral diseases that are common in intensive pig production. No studies compared APP levels for viral or bacterial causes of the same clinical sign, so approximations were made based on the limited literature investigating diseases with different aetiologies. C-reactive protein appears to alter a larger APR in bacterial disease than in viral disease and as a result, CRP may be a candidate for guiding antibiotic initiation in pigs and evaluate treatment efficacy, though additional research is required to confirm this. Furthermore, the interpretation depends on the magnitude of the increase and thus it should be noted that parturition, age, and general health status influence baseline levels.

Horses

Compared to pigs, not many studies were found investigating experimentally induced diseases in horses. However, one study induced ascending placentitis in mares by intracervical inoculation with Streptococcus equi [56]. As can be seen from Table 5, both SAA and HP were significantly increased whereas Fb was not.

Table 5 Overview of acute phase protein (APP) levels in mares for experimentally induced diseases1

Several naturally occurring diseases were investigated e.g. joint disease [57, 58], respiratory disease [59, 60], intestinal disease [59, 61], infection with L. interrogans [62] and parasitic disease [63,64,65]. The investigated APPs and obtained levels in diseased and non-diseased horses have been summarized in Table 6. All studies investigated SAA, which is the major acute phase protein in horses, but data on alterations in Alb, Fb, HP and CRP was also found. Several of the studies determined aetiology and APP responses in infectious and non-infectious causes of disease. For septic arthritis, significantly higher SAA levels were found in horses with bacterial infections compared to culture negative horses, and levels in fungal infections were in between [58]. In a study on joint diseases e.g. osteoarthritis, osteochondrosis, infectious arthritis and infectious tenovaginitis, significant differences were found between healthy and infectious aetiologies. No difference was shown between infectious and non-infectious causes of disease although large numerical differences were seen [66], perhaps due to low sample sizes in the infectious and non-infectious group respectively (n = 7 and 9) [66]. In foals presented with weakness, pneumonia or diarrhoea, significant differences in SAA and Fb were seen between bacterial and non-bacterial infections, indicating that SAA might for some disease complexes be suitable to monitor antibiotic usage in horses although more research is needed to cooperate the findings.

Table 6 Overview of acute phase protein (APP) levels in horses for common diseases and clinical signs1

The clinical relevance of APPs was investigated in two studies [60, 63]. Combining surfactant protein D and SAA and a cut-off of 0.80 (Log) for the sum, inflammatory airway disease (IAD) and control horses could be distinguished with a sensitivity and specificity of 100% and 90%, respectively. Furthermore, surfactant protein D and HP rendered a sensitivity and a specifictiy of 100%, when using a cut-off of 7.70 (Log) [60]. The diagnostic potential of HP and SAA in the parasitic disease habronemosis has been studied. The study found that using a threshold of 0.85 and 11.66 for HP and SAA provided a sensitivity, specificity, and accuracy of 87%/90%, 81%/86% and 87%/94%, respectively [63]. Using HP would correctly classify 84.3% of the horses and for SAA, this would be true for 88.2%. That APP levels can be used to find parasitic infections is ambiguous since no difference in APP levels was found in protozoal myeloencephalitis and Theileria equi infections [64, 65]. Emphasizing that a one-size-fits-all solution, is not possible regarding the diagnostic and prognostic value of APP levels and that more research is needed before APP levels can be used to guide antimicrobial treatment.

Summary—horses

The major acute phase protein in horses is SAA, but CRP, HP and Fb are frequently studied concurrently. In most of the investigated diseases or syndromes SAA, CRP and HP were significantly elevated in diseased versus non-diseased horses, except in parasitic disease where evidence suggests a lack of response. Regarding the use of APP levels to differentiate between infectious and non-infectious causes of disease, more studies have been carried out for horses than pigs, and although solid evidence is still lacking, SAA could be a candidate.

Cattle

For cattle, the most investigated disease is mastitis, both in its clinical and subclinical form as well as when experimentally induced. Studies involving experimentally induced mastitis focused mainly on E. coli mastitis, and the relevant APPs have been summarized in Table 7. As with naturally occurring mastitis, the main APPs in focus were serum and milk HP and amyloid A. However, albumin was also studied [67]. In serum, both SAA and HP were found to be increased significantly in E. coli and S. aureus infections [68,69,70]. Milk HP was not significantly increased after inoculation with S. aureus [69]. The increase in SAA and HP peaked around 2–5 days after inoculation [67,68,69,70] and the pattern was comparable in milk and serum samples [69].

Table 7 Overview of acute phase protein (APP) levels in cows for experimentally induced diseases1

Calves experimentally induced with bovine besnoitiosis experienced significant alterations in HP and Alb from 8 days post-infection [71]. The parasite was inoculated using three different routes and the HP response was highly method dependent with subcutaneous inoculation giving a faster response than intradermal and intravenous inoculation [71]. The kinetics of SAA, milk amyloid A (MAA) and HP seem comparable to what was found in porcine HP and SAA although HP is classified as a major APP in cows and hence a faster increase and decline would be expected [72]. Suggesting that the kinetics and half-life of the APP response might be both disease and species dependent and should be considered when interpreting APP levels in all three species.

Data on naturally occurring mastitis has been summarized in Table 8, where 10 studies investigated the association between APP levels and subclinical mastitis [73,74,75,76,77,78,79,80,81,82] and four investigated the association with clinical mastitis [78, 80, 83, 84]. In subclinical mastitis the serum and MAA, HP and Fb levels were in most studies significantly increased and albumin decreased compared to healthy cows. Furthermore, if more udder quarters were affected, a higher APP level response was seen in MAA, SAA and HP [75]. These findings suggest that APP levels can be useful in identifying cows with subclinical mastitis. The diagnostic potential of Alb, MAA, SAA serum and milk HP based on ROC-curve analysis has been summarized in Table 9. In short, the sensitivity for all APPs except Albumin was > 90% and the specificity ranged from 56.66 to 100% [73, 76, 77, 82]. The large differences in specificity could be due to variations in cut-off values and case definitions, since the studies diagnosed subclinical mastitis based on different parameters. For clinical cases, all investigated APPs were significantly altered in both milk and serum. The sensitivity and specificity based on ROC-curves are summarized in Table 9. In short, the sensitivities were between 79 and 93% and the specificities ranged from 80 to 100%. In mastitis, the APP response seems to be dependent on aetiology [85,86,87] with Gram-negative bacteria yielding a higher response than Gram-positive ones [85, 86]. A few exceptions exists e.g. in MAA where A. pyogenes, environmental Streptococcus and Enterococcus spp. yielded a high response [86, 87]. One study looked at non-infectious and infectious causes of subclinical mastitis. The data showed that cows with subclinical mastitis and positive bacteriology had higher SAA compared to cows with subclinical mastitis but negative bacteriology [82]. Additionally, cows with subclinical mastitis which resolved spontaneously had lower HP levels compared to non-self-resolving cases. Hence, HP analysis could potentially be used to identify S. aureus subclinical mastitis incidences not requiring antibiotic treatment [81].

Table 8 An overview of acute phase protein (APP) levels in cattle for naturally occurring mastitis1
Table 9 Receiver operating curve data calculated for mastitis using different acute-phase proteins (APPs)1

Literature on APP levels in endometritis [88,89,90], urinary tract infections [83], hoof and leg disorders [91,92,93], foot and mouth disease [94, 95], paratuberculosis [96, 97], bovine viral disease [98], infections with L. interrogans [62], bovine besnoitiosis [71], Babesia bigemina [99], respiratory disease [100], subclinical and clinical ketosis [101], traumatic reticuloperitonitis [102], peritonitis [103] and calf diarrhoea [104] was found and summarized in Table 10. As seen in sows, SAA levels and HP levels increase after parturition, but stay elevated for a prolonged period in cows suffering from endometritis and metritis [90]. The exact period in which SAA and HP levels are significantly increased is equivocal since some discrepancy occurs in results and time period studied [88,89,90]. Regardless, it is concluded that SAA and HP could be reliable indicators of clinical endometritis and metritis [88,89,90], used to monitor treatment [89] and HP to evaluate reproductive performance [90]. Additionally, using a cut-off of 0.085 mg/mL for HP gave a sensitivity and specificity of 95% and 57.5%, respectively, between healthy cows and those with moderate endometritis [88]. For SAA a cut-off of 16.25 μg/mL gave a sensitivity and specificity of 95% and 55% between healthy cows and cows with moderate endometritis [88].

Table 10 Overview of acute phase protein (APP) levels in cows for other common diseases than mastitis

One or two studies per disease were found for urinary tract infections (UTIs), foot and mouth disease (FMD), traumatic reticuloperitonitis and bovine viral disease (BVD) and the studies suggest a significant increase in HP and SAA for all four diseases [83, 94, 95, 98, 102], a significant increase in fibrinogen in traumatic reticuloperitonitis, FMD and UTIs [83, 94, 102] and a significant increase in AGP in UTIs [83]. Furthermore, a study suggests a high diagnostic accuracy of Fb, AGP, HP and SAA in UTIs, with AUC being between 0.93 and 0.98 for all APPs and a high prognostic accuracy of SAA and HP (AUC > 0.95) [83]. The association being that a higher HP or SAA level corresponded to poor treatment response after treatment with antibiotics.

Generally speaking, lameness and claw disorders lead to significant increases in HP, SAA [91,92,93], Fb and CRP [92] compared to healthy controls. However, when studying the aetiology behind claw lesions, heel erosions and white line separation seem to provide a lower and non-significant increase in HP and SAA compared to acute laminitis, sole ulcers and digital dermatitis [91]. A finding partly supported by an Estonian study which found that sole ulcers led to increased SAA levels but not HP levels and that digital dermatitis had no significant effect on HP or SAA level [105]. Additionally, when looking at interdigital laminitis, lameness associated with F. necrophorum provided a quantitatively higher level of HP and SAA compared to lameness not associated with F. necrophorum. Due to low sample size in the F. necrophorum positive group (n = 4) a statistical comparison was not computed [93].

Paratuberculosis in cattle is a non-curable enteric disease caused by M. avium spp. paratuberculosis (MAP). The disease is associated with chronic diarrhoea, and compared to healthy controls, MAP-seropositive animals with chronic diarrhoea had significant alterations in albumin levels but not in HP and SAA [96]. The lack of a response in SAA and HP could be due to lower total protein levels in serum caused by the protein-losing enteropathy associated with paratuberculosis [106, 107]. However, when looking at different pathological forms of paratuberculosis in cattle, multifocal and diffuse paucibacillary lesions but not focal and diffuse multibacillary lesions, lead to significantly increased SAA and Hp levels [97]. This suggests a difference in APP pattern between lesion types. Another explanation to a lack of response in HP and SAA in paratuberculosis could be the chronic nature of the disease, since chronic disease has been shown to relate to lower APP levels compared to acute disease in cattle [108].

Investigations on the relationship between APP levels and health status in calves discovered that inflammatory disorders such as respiratory and gastrointestinal disease was connected to elevated APP responses [104, 109,110,111,112,113], and fibrinogen and HP levels mirrored the faecal score with Fb and HP increasing with increased looseness of the faeces, a pattern not seen in SAA [112]. Furthermore, E. coli diarrhoea gave a higher SAA response than coronavirus based diarrhoea [104]. The effect of sepsis on acute phase proteins was reviewed recently, and it concludes that acute phase proteins are unspecific for sepsis in both pigs, horses and cows, but can provide information on disease development and prognosis [25]. The review includes a study on eight calves with significant HP increases in calves with experimentally induced endotoxemia [25], whereas a study on naturally occurring systemic inflammatory response syndrome in 102 calves showed no increase in HP compared to non-SIRS calves [113].

Summary—cattle

The major acute phase proteins in cattle are serum or milk amyloid A and HP, and the majority of literature reports oncases of clinical and subclinical mastitis. Fb is, as for horses, commonly studied and is significantly lower in diseases such as subclinical and clinical mastitis, urinary tract infections and lameness, which could be valuable for the purposes of diagnostics and treatment. Mastitis, both experimentally induced and field cases, resulted in significant changes in serum and milk amyloid A and HP, and Gram-negative bacteria appeared to produce a higher APP response than Gram-positive bacteria, with few exceptions. Significant increases in serum and milk amyloid A and HP are seen in subclinical mastitis, and the level corresponds to the number of udder quarters affected. Using serum and milk HP and amyloid A as a diagnostic marker of subclinical mastitis yielded a sensitivity of more than 90% and a specificity ranging from 56.66 to 100% depending on the cut-off used. Furthermore, one study suggested that SAA could be used to differentiate between subclinical cases with and without positive bacteriology, and HP could be used to differentiate S. aureus cases that will self-resolve. For some other diseases than mastitis, HP and SAA are elevated, except for besnoitiosis, some claw lesions and paratuberculosis where evidence is equivocal. Furthermore, studies suggested that SAA and HP could be used to monitor and predict treatment response in uterine tract infections and endometritis and that lameness associated with F. necrophorum infections causes higher HP and SAA responses than non-F. necrophorum interdigital laminitis. In calves, diarrhoea of bacterial origin yielded a higher SAA response than viral diarrhoea.

Conclusions

Much research has been put into identifying which APPs are elevated or decreased in diseases affecting, pigs, horses and cows and correlating clinical signs with the APR. Generally, most diseases and clinical signs of disease initiate an APR in the moderate to major APPs for all species. Few studies, however, evaluated the diagnostic and prognostic potential. Research on the APR in humans has shown that APPs can be used diagnostically e.g., to comprehend the seriousness of a condition, to distinguish infectious from non-infectious causes of disease, or bacterial from viral causes or guide antibiotic treatment initiation and efficacy. This review found limited evidence within pigs, horses, and cattle on either of these subjects.

Eleven studies used ROC-curve analysis to establish the diagnostic performance of an APP in differentiating diseased from non-diseased animals. Six of the eleven studies targeted clinical or subclinical mastitis. Using APP levels to distinguish between healthy and sick animals is valuable in conditions where welfare or production outcomes are affected before clinical signs arise as is the case with subclinical mastitis. However, the value is limited in diseases such as lumpy skin disease where the diagnosis can be determined from clinical signs alone or in multifactorial disease where both bacterial, viral and non-infectious causes lead to similar clinical signs e.g., diarrhoea or lameness. In the latter cases quantification of the acute phase level is valuable for the veterinarian or the farmer if it can be correlated to seriousness of the disease, prognosis, aetiology, or treatment.

This review identified 10 studies relating acute phase protein levels to aetiology, three to antibiotic treatment efficacy, one to prognosis and two to seriousness of disease. However, these studies concern different species and diseases, hence it is difficult to definitively infer upon the usefulness of APPs to mitigate unnecessary antibiotic use. Nevertheless, with the limited amount of evidence in mind, CRP might have the potential to differentiate between bacterial and non-bacterial causes of disease in pigs. SAA could reflect underlying aetiology in horses, and in mastitis and neonatal diarrhoea in cattle.

Since reducing antibiotic usage is of high priority, tools to guide initiation and termination would be of great interest. Identification of APPs could aid in separating bacterial from viral or non-infectious causes of disease, predict prognosis and be used to evaluate antibiotic treatment efficacy. More evidence should be collected for each disease within a species where mitigation of antibiotic usage is relevant and underlying aetiology cannot be determined by the clinical signs alone or where pre-treatment diagnostics are not viable. It is the authors’ belief that to determine the viability of using APPs to mitigate unnecessary antibiotic use, more efforts should be put into correlating APP levels and underlying aetiology, establishing cut-off values for infectious and non-infectious cases, evaluate the diagnostic performance of APPs and evaluate any adverse effects of using APPs to determine treatment strategy.